What a finance audit finds in the first two weeks

We start every engagement with a two-week audit of the finance function: the model, the close, the systems, the metric definitions. It is meant to be uncomfortable, and the striking thing is how little the findings vary between companies.

First, the same metric has more than one definition in active use. ARR in the board deck does not match ARR in the sales dashboard, usually because one includes services revenue or counts contracts at signature rather than start. Nobody is wrong; nobody wrote it down.

Second, the model has hardcoded values in the forecast period. Almost always. Frequently the person who put them there has left, and nobody remaining is willing to change them because nobody knows what they were compensating for.

Third, the chart of accounts has grown without pruning. Departments that no longer exist, accounts created for a single transaction in 2023, cost centres that map to nothing. This is why departmental reporting takes four days and why the numbers shift depending on who pulls them.

Fourth, the close has no written owner per task. Everyone knows roughly what they do; nobody knows what they are blocking. This is the single largest driver of close duration and the cheapest to fix.

Fifth, cohort data exists but has never been assembled. The raw information is sitting in the billing system, and no one has built the retention curve, which means pricing and expansion decisions are being made on aggregate averages that hide everything interesting.

None of these are failures of competence. They are what happens when a finance function grows by accretion while the company is busy doing something more urgent. They are also all fixable in a quarter, which is the more useful half of the finding.